Optimum Receiver Design and Performance Analysis of Arbitrarily Correlated Rician Fading MIMO Channels With Imperfect Channel State Information.

The error performance of different types of receivers over an arbitrarily correlated Rician fading MIMO channel is assessed by evaluating the decision metric and the pairwise error probability. The differences are in the way channel state information at the receiver (CSIR) is recovered. We consider: i) a genie-aided receiver, based on the perfect knowledge of CSIR; ii) a mismatched receiver, based on the use of pilot-aided (imperfect) CSIR in the perfect-CSIR decision metric; iii) an optimum receiver, based on perfect channel distribution information at the receiver (CDIR), which performs jointly channel and data estimation. In this paper, the decision metric of the optimum receiver is derived, and an iterative algorithm is proposed to calculate it for sequential decoding schemes. The complexity of this iterative algorithm is analyzed and compare against the complexity of the mismatched receiver. Closed-form expressions are given for the pairwise error probabilities (PEPs) of the three receivers with arbitrarily correlated Rician fading. These expressions are further processed to obtain the asymptotic (in the SNR) PEPs, the diversity order, and the asymptotic power losses. Numerical results are presented to support the validity of this analysis and to assess the impact of power efficiency and CDIR estimation errors on the PEP versus the εb/N0 ratio. [ABSTRACT FROM AUTHOR]